Openai Releases Math And Computer Science Claims, Holds Back Physics Work
OpenAI has published AI-generated results in mathematics and computer science while asking outside experts to examine those claims and guide the release of separate physics work.
OpenAI has announced AI-generated research results in mathematics, computer science and physics, following its recent release of hundreds of manuscripts on math problems. This time, the company is using different approaches to sharing the work: the math and computer science materials are public, while an external group is overseeing the unreleased physics materials. The claims still need to be examined and explained by researchers.
Table of Contents
Two results are available for scrutiny
The published math manuscript presents a proof of Sogge’s local smoothing for waves in every spatial dimension of at least two, according to The Verge. The outlet reports that OpenAI says the proof also resolves the Kakeya maximal and dimension conjectures and yields Stein’s spherical restriction estimates within the stated range. For this result, OpenAI released the model output directly so researchers can assess it and develop their own explanations.
The computer science manuscript makes a more surprising claim: that the Exponential Time Hypothesis, or ETH, is false. The Verge says the proposed proof gives a deterministic algorithm for exactly counting solutions to certain fixed-width formulas in subexponential time and polynomial space. Engadget notes that ETH is widely believed and underlies hundreds of papers, including work in databases and computational biology. If the claim holds up, it could point toward more efficient algorithms for some problems in those fields.
Outside researchers will examine the claims
OpenAI has asked a working group led by Russell Impagliazzo, Ramamohan Paturi and Ryan Williams to examine and explain the computer science result, according to The Verge. Impagliazzo and Paturi helped propose ETH. The group requested that the AI output be made public immediately so others could examine it too; The Verge reports that a Lean proof was released alongside the write-up.
The physics claim concerns a rigorous existence proof for interacting quantum field theories with gauge fields and fermions in four dimensions. Unlike the other two results, the full physics materials have not been released. The Verge reports that a group led by David Gross will guide that work; Engadget says OpenAI has left decisions about its release to the group.
Publication and credit remain open questions
Engadget reports that OpenAI does not plan to seek academic credit for the models’ discoveries. Instead, the external groups would receive credit for their own write-ups and materials and would be responsible for appropriately crediting the researchers and ideas on which the AI-generated work draws.
The approach has prompted questions about who gets early access to AI-generated results. According to The Verge, OpenAI’s AGMAI panel of math advisers recommended against giving selected people the output first, citing concerns including unequal access and biased selection. The computer science group has also asked the research community to discuss access, conflicts of interest and how to credit the work of checking and explaining results. Those questions matter for researchers and students trying to evaluate new claims while maintaining the expertise to challenge them.
Sources
This story was compiled by AI from the reports below. Read the originals for the full details.